Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Nonlocal estimation of manifold structure.

Yoshua Bengio1, Martin Monperrus, Hugo Larochelle

  • 1bengioy@iro.umontreal.ca

Neural Computation
|August 16, 2006
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Imagining and building wise machines: the centrality of AI metacognition.

Trends in cognitive sciences·2026
Same author

Navigating Ternary Doping in Li-ion Cathodes With Closed-Loop Multi-Objective Bayesian Optimization.

Advanced materials (Deerfield Beach, Fla.)·2026
Same author

Divergent creativity in humans and large language models.

Scientific reports·2026
Same author

Identifying indicators of consciousness in AI systems.

Trends in cognitive sciences·2025
Same author

Publisher Correction: Deep-learning-based virtual screening of antibacterial compounds.

Nature biotechnology·2025
Same author

Deep-learning-based virtual screening of antibacterial compounds.

Nature biotechnology·2025

Local manifold learning algorithms struggle with high-dimensional data. Nonlocal methods, by analyzing tangent planes across different data points, offer improved generalization, especially in complex scenarios like image rotation learning.

Area of Science:

  • Machine Learning
  • Dimensionality Reduction
  • Computer Vision

Background:

  • Many manifold learning algorithms are local, kernel-based methods.
  • These local methods are susceptible to the curse of dimensionality on the underlying manifold.

Purpose of the Study:

  • To explore nonlocal manifold learning algorithms.
  • To propose a training criterion for nonlocal methods.
  • To demonstrate the advantages of nonlocal approaches over local ones.

Main Methods:

  • Framing local manifold learning as kernel learning.
  • Developing a training criterion for nonlocal manifold learning.
  • Experimentally evaluating a tangent plane prediction function.

Main Results:

Related Experiment Videos

  • Local manifold learning algorithms suffer from the curse of dimensionality.
  • Nonlocal algorithms can discover shared structure in tangent planes.
  • The proposed nonlocal method generalizes effectively to unseen data, outperforming local methods.

Conclusions:

  • Nonlocal manifold learning algorithms are a promising alternative to local methods.
  • These algorithms overcome limitations imposed by the curse of dimensionality.
  • The tangent plane approach offers significant advantages in generalization capabilities.